Samuel Hornus wrote:

> Well, I started the 'fully by hand' method. Booting 'boot.fs' without
> any problem.
> 
> BUT :
> 
> My primary HD (wd0) is 'master' and have MacOS9.1 and MacOSX on it. I
> don't want to touch it.
> My secondary HD (wd1 - is it really wd1 or am I mistaking ?) is set to
> 'slave'.
> 
> In order to know the partition scheme that is seen by NetBSD, I type :
> 
> disklabel wd1
> 
> It works fine with 'disklabel wd0' but fails with wd1 (that would be too
> easy ;-))
> The error is :  'wdc1 (or wd1c) lost interrupt'
> 
> So I can't continue the intall since I can't even blindly newfs wd1
> partitions... :-(

OK.  Do you want a mixed config on the second drive, or NetBSD only?  If you
want NetBSD only, then you can simply clobber the drive with sysinst and go
from there.

If you want a mixed config, you need to follow the instructions I gave (or
similar).


As to the wd0 vs wd1 issue, the easiest way to fix this (assuming two IDE
drives on a single bus) is to simply open the case and swap the master and
slave jumpers on the two drives.  This also makes the old slave the primary
(wd0) and the old primary the slave (wd1).  MacOS doesn't care, and NetBSD
is happy.


BTW: Have you partitioned the secondary drive at all?


> 1: Does anyone has an idea the error come from ?
> 2: what is the difference between /dev/wd1a and /dev/rwd1a ?

Not sure off the top of my head.  I think the wd vs rwd is a character
special vs block special issue (two different ways of viewing and accessing
the same physical device), but I warn that that is a guess from memory.  I
haven't checked this against the documentation.

-- 
Andrew White                Andrew.E.White@motorola.com